Research On Electrical And Control System And Object Recognition Algorithm For Water Area Garbage Cleaning Robot

With the rapid development of China's comprehensive national strength,there are a large number of various garbage and waste produced in people's daily life and production.Many water environments are often threatened by floating garbage,which seriously affects the water landscape and water quality,resulting in water body stinking,silting,and even destroying the diversity and ecological balance of aquatic species.Environmental pollution has also awakened people's awareness of environmental protection and governance.At present,there are various large-scale cleaning vessels with high degree of mechanization and high efficiency.However,they cannot be used in small waters due to their complicated structure,large size,poor flexibility and high cost.Floating garbage in small waters mainly relies on manual salvage,which has disadvantages such as high labor consumption,high risk coefficient and low efficiency.Therefore,this paper studied and designed a small water area garbage cleaning robot based on intelligence,high efficiency and environmental protection,and premised on low cost and large-scale promotion.The main research contents include:(1)Design of the overall scheme and mechanical structureBased on the analysis of the task requirements and functions of the small water area garbage cleaning robot,two cleaning modes of intelligent identification and remote control was proposed.The framework design of the whole system was completed.The mechanical structure design requirements of the robot are proposed.Solidworks was used for 3D modeling of mechanical structure,including garbage collection device,propulsion element,collection network,shell and buoy,etc.The parts of the robot were processed and assembled,and the prototype of the robot was made finally.(2)Design and research on power supply system based on solar energyIn order to make the robot energy self-supporting,a solar-based power supply system was designed.By analyzing the running resistance of the robot in water,the power demand of the robot was calculated.On this basis,the corresponding power supply system scheme was designed,and the reasonable power supply system capacity parameters are designed according to the working requirements of the robot and the local illumination radiation quantity.In order to make the system output higher quality,more powerful energy,solar battery of Simulink model was established to analyze the output characteristics under different environments.An improved adaptive MPPT algorithm was proposed by analyzing the anvantages and disadvantages of classical maximum power point tracking algorithm,and the simulation model was built on Simulink for comparative experiments.The simulation results show that the proposed method can achieve the good results of short control time,low oscillation and low power loss.(3)Research on water surface object recognition algorithmIn order to realize the intelligent operation of the robot,computer vision was applied to the robot surface recognition technology,and the identification algorithm of common garbage and obstacles on the water surface was studied.By analyzing the difficulty of water surface image processing,the improved Meanshift+OTSU method was proposed to achieve water surface target detection.This method can overcome the influence of water surface ripple and uneven illumination to some extent.Taking plastic bottles,fallen leaves,stones and lotus leaves as the recognition targets,the color moment,gray level co-occurrence matrix moment and HU invariant moment were extracted.The Relief F algorithm was used to select and fuse features,which can better represent the original properties of specific objects.The SVM was used to establish the recognition model.Taking accuracy rate,false detection rate and missing detection rate as performance indexes,the effects of different feature combinations on the recognition of targets were compared by experiments.The results show that the proposed method exhibits superior recognition performance.(4)Design and implementation of control systemThe robot control system based on Rasphberry Pi 3B micro controlller was designed,and Raspbian system was adopted as the software platform.Various robot functional tests were performed on the artificial lake.The test content includes: power supply performance test,garbage collection function test,intelligent identification function test and remote control function test.The test results show that the solar power supply system can meet the normal power demand of the robot,which solves the problem of endurance to a certain extent.The robot can run smoothly on the artificial lake,and its garbage collection device can collect garbage efficiently and reliably.Through the cooperation of intelligent recognition and remote control,the robot can clean the common floating garbage intelligently and automatically.The research of this subject provides solutions and theoretical support for the small-scale water garbage cleaning device.The development of the robot provides convenience for the cleaning of floating garbage,which has certain using significance and promotion value.